EE313 Linear Systems and Signals

Fall 2010

Sampling Theorem
Prof. Brian L. Evans
Dept. of Electrical and Computer Engineering
The University of Texas at Austin
Initial conversion of content to PowerPoint
by Dr. Wade C. Schwartzkopf

Sampling: Time Domain

Many signals originate as continuous-time
signals, e.g. conventional music or voice
By sampling a continuous-time signal at
isolated, equally-spaced points in time, we
obtain a sequence of numbers s
t
s n s n Ts

sampled

Ts

n {, -2, -1, 0, 1, 2,}

Ts is the sampling period.

t
Ts

s(t)

ssampled t s (t ) t n Ts
n

impulse train

Sampled analog waveform

16 - 2

Sampling: Frequency Domain

Replicates spectrum of continuous-time signal
At offsets that are integer multiples of sampling frequency

Fourier series of impulse train where s = 2 fs

1 2
2
T (t ) t n Ts cos( s t ) cos(2 s t ) . . .
s

Ts

g (t ) f (t ) Ts (t )

Ts

Ts

1
f (t ) 2 f (t ) cos(s t ) 2 f (t ) cos(2 s t ) . . .
Ts
Modulation by cos( s t)

F()
Example

G()

-2fmax 2fmax

Modulation by cos(2 s t)

gap if and only if 2f max 2f s 2f max f s 2 f max 16 - 3

Shannon Sampling Theorem

A continuous-time signal x(t) with frequencies no
higher than fmax can be reconstructed from its
samples x[n] = x(n Ts) if the samples are taken at
a rate fs which is greater than 2 fmax.
Nyquist rate = 2 fmax
Nyquist frequency = fs/2.

What happens if fs = 2fmax?

Consider a sinusoid sin(2 fmax t)
Use a sampling period of Ts = 1/fs = 1/2fmax.
Sketch: sinusoid with zeros at t = 0, 1/2fmax, 1/fmax,

16 - 4

Shannon Sampling Theorem

Assumption
Continuous-time signal has
no frequency content above
fmax
Sampling time is exactly the
same between any two
samples
Sequence of numbers
obtained by sampling is
represented in exact
precision
Conversion of sequence to
continuous time is ideal

In Practice

16 - 5

Why 44.1 kHz for Audio CDs?

Sound is audible in 20 Hz to 20 kHz range:
fmax = 20 kHz and the Nyquist rate 2 fmax = 40 kHz

What is the extra 10% of the bandwidth used?

Rolloff from passband to stopband in the magnitude
response of the anti-aliasing filter

Okay, 44 kHz makes sense. Why 44.1 kHz?

At the time the choice was made, only recorders capable of
storing such high rates were VCRs.
NTSC: 490 lines/frame, 3 samples/line, 30 frames/s =
44100 samples/s
PAL: 588 lines/frame, 3 samples/line, 25 frames/s = 44100
samples/s
16 - 6

Sampling
As sampling rate increases, sampled waveform
looks more and more like the original
Many applications (e.g. communication
systems) care more about frequency content in
the waveform and not its shape
Zero crossings: frequency content of a sinusoid
Distance between two zero crossings: one half period.
With the sampling theorem satisfied, sampled sinusoid
crosses zero at the right times even though its
waveform shape may be difficult to recognize
16 - 7

Aliasing
Analog sinusoid
x(t) = A cos(2f0t + )

Sample at Ts = 1/fs
x[n] = x(Ts n) =
A cos(2f0 Ts n + )

Keeping the sampling

period same, sample

y[n] = y(Ts n)
= A cos(2(f0 + lfs)Tsn + )
= A cos(2f0Tsn + 2lfsTsn + )
= A cos(2f0Tsn + 2l n + )
= A cos(2f0Tsn + )
= x[n]

Here, fsTs = 1
Since l is an integer,
cos(x + 2l) = cos(x)

y(t) = A cos(2(f0 + lfs)t + ) y[n] indistinguishable

where l is an integer

from x[n]

Frequencies f0 + l fs for l 0 are aliases of frequency f0 16 - 8